File Download
 
Links for fulltext
(May Require Subscription)
 
Supplementary

Article: Bilinear probabilistic principal component analysis
  • Basic View
  • Metadata View
  • XML View
TitleBilinear probabilistic principal component analysis
 
AuthorsZhao, J
Yu, PLH
Kwok, JT
 
KeywordsBilinear systems
Data reduction
Parameter estimation
Principal component analysis
Probability
 
Issue Date2012
 
PublisherIEEE. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=72
 
CitationIEEE Transactions on Neural Networks and Learning Systems, 2012, v. 23 n. 3, p. 492-503 [How to Cite?]
DOI: http://dx.doi.org/10.1109/TNNLS.2012.2183006
 
AbstractProbabilistic principal component analysis (PPCA) is a popular linear latent variable model for performing dimension reduction on 1-D data in a probabilistic manner. However, when used on 2-D data such as images, PPCA suffers from the curse of dimensionality due to the subsequently large number of model parameters. To overcome this problem, we propose in this paper a novel probabilistic model on 2-D data called bilinear PPCA (BPPCA). This allows the establishment of a closer tie between BPPCA and its nonprobabilistic counterpart. Moreover, two efficient parameter estimation algorithms for fitting BPPCA are also developed. Experiments on a number of 2-D synthetic and real-world data sets show that BPPCA is more accurate than existing probabilistic and nonprobabilistic dimension reduction methods.
 
ISSN2162-237X
2012 Impact Factor: 3.766
 
DOIhttp://dx.doi.org/10.1109/TNNLS.2012.2183006
 
ISI Accession Number IDWOS:000302705100010
 
DC FieldValue
dc.contributor.authorZhao, J
 
dc.contributor.authorYu, PLH
 
dc.contributor.authorKwok, JT
 
dc.date.accessioned2012-04-24T07:52:49Z
 
dc.date.available2012-04-24T07:52:49Z
 
dc.date.issued2012
 
dc.description.abstractProbabilistic principal component analysis (PPCA) is a popular linear latent variable model for performing dimension reduction on 1-D data in a probabilistic manner. However, when used on 2-D data such as images, PPCA suffers from the curse of dimensionality due to the subsequently large number of model parameters. To overcome this problem, we propose in this paper a novel probabilistic model on 2-D data called bilinear PPCA (BPPCA). This allows the establishment of a closer tie between BPPCA and its nonprobabilistic counterpart. Moreover, two efficient parameter estimation algorithms for fitting BPPCA are also developed. Experiments on a number of 2-D synthetic and real-world data sets show that BPPCA is more accurate than existing probabilistic and nonprobabilistic dimension reduction methods.
 
dc.description.naturepublished_or_final_version
 
dc.identifier.citationIEEE Transactions on Neural Networks and Learning Systems, 2012, v. 23 n. 3, p. 492-503 [How to Cite?]
DOI: http://dx.doi.org/10.1109/TNNLS.2012.2183006
 
dc.identifier.doihttp://dx.doi.org/10.1109/TNNLS.2012.2183006
 
dc.identifier.epage503
 
dc.identifier.hkuros199326
 
dc.identifier.isiWOS:000302705100010
 
dc.identifier.issn2162-237X
2012 Impact Factor: 3.766
 
dc.identifier.issue3
 
dc.identifier.spage492
 
dc.identifier.urihttp://hdl.handle.net/10722/146419
 
dc.identifier.volume23
 
dc.languageeng
 
dc.publisherIEEE. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=72
 
dc.publisher.placeUnited States
 
dc.relation.ispartofIEEE Transactions on Neural Networks and Learning Systems
 
dc.rightsIEEE Transactions on Neural Networks and Learning Systems. Copyright © IEEE.
 
dc.rights©2012 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
 
dc.rightsCreative Commons: Attribution 3.0 Hong Kong License
 
dc.subjectBilinear systems
 
dc.subjectData reduction
 
dc.subjectParameter estimation
 
dc.subjectPrincipal component analysis
 
dc.subjectProbability
 
dc.titleBilinear probabilistic principal component analysis
 
dc.typeArticle
 
<?xml encoding="utf-8" version="1.0"?>
<item><contributor.author>Zhao, J</contributor.author>
<contributor.author>Yu, PLH</contributor.author>
<contributor.author>Kwok, JT</contributor.author>
<date.accessioned>2012-04-24T07:52:49Z</date.accessioned>
<date.available>2012-04-24T07:52:49Z</date.available>
<date.issued>2012</date.issued>
<identifier.citation>IEEE Transactions on Neural Networks and Learning Systems, 2012, v. 23 n. 3, p. 492-503</identifier.citation>
<identifier.issn>2162-237X</identifier.issn>
<identifier.uri>http://hdl.handle.net/10722/146419</identifier.uri>
<description.abstract>Probabilistic principal component analysis (PPCA) is a popular linear latent variable model for performing dimension reduction on 1-D data in a probabilistic manner. However, when used on 2-D data such as images, PPCA suffers from the curse of dimensionality due to the subsequently large number of model parameters. To overcome this problem, we propose in this paper a novel probabilistic model on 2-D data called bilinear PPCA (BPPCA). This allows the establishment of a closer tie between BPPCA and its nonprobabilistic counterpart. Moreover, two efficient parameter estimation algorithms for fitting BPPCA are also developed. Experiments on a number of 2-D synthetic and real-world data sets show that BPPCA is more accurate than existing probabilistic and nonprobabilistic dimension reduction methods.</description.abstract>
<language>eng</language>
<publisher>IEEE. The Journal&apos;s web site is located at http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=72</publisher>
<relation.ispartof>IEEE Transactions on Neural Networks and Learning Systems</relation.ispartof>
<rights>IEEE Transactions on Neural Networks and Learning Systems. Copyright &#169; IEEE.</rights>
<rights>&#169;2012 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.</rights>
<rights>Creative Commons: Attribution 3.0 Hong Kong License</rights>
<subject>Bilinear systems</subject>
<subject>Data reduction</subject>
<subject>Parameter estimation</subject>
<subject>Principal component analysis</subject>
<subject>Probability</subject>
<title>Bilinear probabilistic principal component analysis</title>
<type>Article</type>
<description.nature>published_or_final_version</description.nature>
<identifier.doi>10.1109/TNNLS.2012.2183006</identifier.doi>
<identifier.hkuros>199326</identifier.hkuros>
<identifier.volume>23</identifier.volume>
<identifier.issue>3</identifier.issue>
<identifier.spage>492</identifier.spage>
<identifier.epage>503</identifier.epage>
<identifier.isi>WOS:000302705100010</identifier.isi>
<publisher.place>United States</publisher.place>
<bitstream.url>http://hub.hku.hk/bitstream/10722/146419/1/Content.pdf</bitstream.url>
</item>